Author: Matt Corallo 2021-08-16 04:48:36
Published on: 2021-08-16T04:48:36+00:00
In this email thread, Anthony Towns and Matt Corallo discuss the issue of base fees in lightning routing algorithms. Anthony argues that base fees make the algorithm slower while Matt argues that base fees are necessary to match the costs incurred by nodes for forwarding an HTLC. However, Anthony disagrees with Matt's point and says that base+proportional fees paid only on success roughly match the value of forwarding an HTLC but not the costs. The costs are incurred on failed HTLCs as well and depend on the time a HTLC lasts, and also vary heavily depending on how many other simultaneous HTLCs there are. They further discuss the opportunity cost of not being able to accept other transactions that would pay higher fees, which can be captured equally well by just setting a proportional fee and a minimum payment value.Matt says that he doesn't believe upfront fees and refunds on HTLC fulfillment are the right approach either, but no one will realize it until he implements his solution. He thinks that these approaches may be better than any other proposal currently available but not particularly good. They then discuss the cost to nodes, which is the opportunity cost of not being able to accept other transactions that would pay higher fees. Matt argues that setting a proportional fee captures this cost equally well, which Anthony disagrees with. Anthony says that you could amortize it, but that doesn't make it "equally" good. He thinks that it's absurd to suggest that this is a perfectly fine thing to do absent a somewhat compelling reason.They also discuss the issue of HTLC slot usage after switching to eltoo. Matt argues that the HTLC slot limit is to keep transactions broadcastable and it won't change even after switching to eltoo. Anthony disagrees and says that eltoo gives us the ability to have channel factories. Finally, they discuss the possibility of dropping base fees entirely and what would happen if over 50% of the network deliberately switched to zero base fees and found it worked fine. Matt suggests that software vendors wishing to rely on these types of algorithms should do the legwork to see what other ideas can be explored before jumping to "ignore all the nodes who've decided their fees are X" because he thinks that's a bad idea for the long-term health of the network.
Updated on: 2023-05-23T15:45:39.578935+00:00